**** BEGIN LOGGING AT Fri Apr 15 02:59:58 2016 Apr 15 08:48:47 hi Apr 15 08:49:41 does anybody have any info on what's happening with oesf.org? It's been offline since two days... does anybody know any of its admins? Apr 15 08:53:31 TheWalt has an account on the archlinuxarm forum. He used to be an admin I thought. Apr 15 08:54:05 thanks daal, I'll try to contact him there Apr 15 08:54:20 in the meantime we might just use https://archlinuxarm.org/forum/viewtopic.php?f=27&t=9381 Apr 15 08:54:27 I hope OESF comes back online Apr 15 08:54:43 yes, good idea Apr 15 08:55:24 and I hope that too, I wanted to post stuff there related to the Zaurus :/ Apr 15 08:57:27 btw is TheWalt = danboid on OESF? Apr 15 08:58:30 no Apr 15 09:08:00 there is an email contact here too : https://whois.domaintools.com/oesf.org Apr 15 09:09:45 nice find, I'll send an email to that one too Apr 15 09:10:22 hello Apr 15 09:10:53 also danboid has an github account https://github.com/danboid/ZALARM-install Apr 15 09:11:01 we will keep it going I guess Apr 15 09:11:10 Hi ant_work Apr 15 09:12:25 daal: I'm busy on two fronts: sound for 4.4 and new issues with 4.6 (doesn't finsih boot) Apr 15 09:12:40 as for 4.5...well, it lacks the keyboard :/ Apr 15 09:12:51 ant_work : I would like to stick with 4.4 atm Apr 15 09:13:13 about sound I think there is a little bug somewhere Apr 15 09:13:14 any progress welcome Apr 15 09:13:29 well, if compiled into kernel, sound works fine Apr 15 09:13:30 I get working alsa if I do modprobe snd_soc_pxa2xx Apr 15 09:13:37 the rest is loaded by udev Apr 15 09:13:55 that is acceptable Apr 15 09:14:07 I'm trying to understand what's wrong on kernel boot: it could be a missing clock Apr 15 09:14:16 hi ant_work Apr 15 09:15:08 ant_work : I had no success with the screen blank issue on 4.4 so far, any news ? Apr 15 09:15:33 btw : OESF is down Apr 15 09:15:48 let me remember, it stays blank after resume? Apr 15 09:16:21 no, first blank goes white, all further blanks are proper black /off Apr 15 09:16:40 if you add a blank during boot that fixes the issue, but dirty workaround Apr 15 09:17:05 probably happens with pxa-fb, on corgi (C860) I don't see it (iirc) Apr 15 09:17:11 I'll test again Apr 15 09:17:35 it was introduced with the 4.x series Apr 15 09:19:21 fwiw see about sound: http://pastebin.com/nkkDynBT Apr 15 09:19:26 anyone has a recent OSEF dump ? Apr 15 09:20:01 failed: -517 seems the driver is returning -EPROBE_DEFER Apr 15 09:20:13 where does this Apr 15 09:20:14 come from : is it a request_gpio(), clk_get(), request_irq(), ... ? Apr 15 09:20:52 my suspects are for clk_get() but I have still to ftrace Apr 15 09:22:21 the issue is for corgi too? Apr 15 09:22:54 211111111111111111111111111111111111111111111--------------------------------- Apr 15 09:23:04 cat on the keyboard :) Apr 15 09:24:19 yes, iirc spitz/akita just have a different codec (WM8750) Apr 15 09:25:45 I see, Apr 15 09:26:15 may help to compare code from 3.1x to 4.x to see what changed ? Apr 15 09:26:32 I am happy with inbuild sound support Apr 15 09:26:54 but there is a bug somewhere Apr 15 09:28:17 happy to test on borzoi Apr 15 09:28:34 I think 4.4 is LTS, best to stay with that Apr 15 09:31:28 ant_work : what was the difference between C1000 and C3x00 again ? My kexecboot does not work on C1000 for some reason Apr 15 09:32:18 there was a reason why different images were provided for these platforms Apr 15 09:32:19 there is a gpio expander, pca953 Apr 15 09:32:38 so why not a uniform kernel Cxx00 ? Apr 15 09:33:02 because the mtdparts are different Apr 15 09:33:31 so this stops the kernel to boot ? I see Apr 15 09:34:08 so I need to compile for C1000 only leaving C3x00 unchecked and enable pca953 ? Apr 15 09:34:31 try just enabling that, it should be ignored by spitz Apr 15 09:34:41 ok Apr 15 09:35:49 thanks for your help daal, I'm looking forward for a new kexecboot for the 1000 Apr 15 09:36:49 VartiWork, I did manage to rebuild the 3.2 for all Zaurus but somehow they don't boot anymore on my devices Apr 15 09:37:09 I bet with gcc5 I need to use more modern kernels Apr 15 09:38:28 can you rebuild on downgraded gcc ? (4.8 ?) Apr 15 09:38:33 daal, the subtle thing is, the Z bootloader is obsolete and the kernel cannot read the atag list (mtdparts included) Apr 15 09:38:55 so the mtdparts are hardcoded in spitz.c and corgi.c Apr 15 09:39:09 but spitz means: 3000, 3100 and 3200 Apr 15 09:39:21 (spitz, borzoi, terrier) Apr 15 09:39:33 kernel only knows about spitz and borzoi Apr 15 09:39:51 thus for 3200 a patch is needed Apr 15 09:40:10 and for c1000 ? Apr 15 09:40:24 different size again :) they had fun isn't? Apr 15 09:41:06 yeah not sure what the sharp linux team was smoking Apr 15 09:41:38 heh, good question Apr 15 09:41:53 my kexec is based on 3.5, so your last was 3.2 ? Apr 15 09:42:26 yes, It was the one I released long ago Apr 15 09:42:48 on my devices I have probably 3.17/18 can't remember. Apr 15 09:42:54 ok, I might use an older archlinuxarm toolchain with gcc 4.8 on a 3.2 kernel for C1000 Apr 15 09:43:38 you still have some of the 3.2 binaries ? Apr 15 09:44:18 I remember after 3.5 the kernel got too big even with lzma or xz Apr 15 09:44:25 not the old ones, have been lost with kexecboot.org outage Apr 15 09:44:34 bugger Apr 15 09:44:48 and the new rebuilt ones don't boot... Apr 15 09:44:55 same kernel + patchset Apr 15 09:44:59 have you tried with older gcc ? Apr 15 09:45:18 no, heh..this would be complicated for me Apr 15 09:45:57 I know you use OE Apr 15 09:45:59 but I'm sure the only device with size problem was collie Apr 15 09:46:09 can you get me the source of your 3.2 ? Apr 15 09:46:20 I can backport gcc toolchain easy Apr 15 09:47:07 I had issues with size but my kexec was not klibc I think Apr 15 09:47:14 what about musl ? Apr 15 09:50:07 musl is implemented by I have yet to build an image with it Apr 15 09:52:05 daal: I can send you/upload later the sources Apr 15 09:52:16 ok great Apr 15 09:52:28 I am busy atm but will give it a go next week Apr 15 09:52:29 for sure I used those sources in 2012 when I uploaded on kexecboot.org Apr 15 09:52:32 http://cgit.openembedded.org/meta-handheld/commit/recipes-kernel/linux?id=2be1fb258598a7e5ef231377f5f9b99d0c2e17ed Apr 15 09:52:53 no patches Apr 15 09:53:34 bluelightning: hi Apr 15 09:53:40 can you provide the kexec initramfs, would save me some time Apr 15 09:54:09 oh, sure Apr 15 09:54:28 github ? Apr 15 09:55:26 but I could inject my latest initramfs based on kexecboot 6.0 ? any C1000 specific changes needed ? Apr 15 09:56:03 ext4 ? Apr 15 09:56:10 there was a little diff... Apr 15 09:56:16 http://cgit.openembedded.org/meta-handheld/tree/recipes-bsp/kexecboot/kexecboot-handhelds.inc Apr 15 09:56:27 spitz needs 1-2 seconds more to scan hdd Apr 15 09:57:22 3 seconds is sometimes not enough with old/big MMC/SD cards Apr 15 09:57:56 well, you just do 'rescan' in those cases Apr 15 09:58:22 apart this the binary is the same Apr 15 10:04:56 thanks Apr 15 10:13:54 ant_work : I am testing a alpine-linux base build server atm, I might give it a go to compile a kexec / base kernel with it\ Apr 15 10:16:54 *musl Apr 15 11:00:50 daal: afaik kexec-tools should compile just fine against musl Apr 15 11:03:05 ah, you maybe need a couple of patches Apr 15 11:06:38 i.e. kexec-tools: Define _GNU_SOURCE for getting loff_t definition **** ENDING LOGGING AT Sat Apr 16 02:59:58 2016